The Corner

Politics & Policy

Two Mistakes to Avoid on Trump Polling

An NBC/Marist poll released earlier this week found that President Trump has “dismal job approval ratings” in Michigan, Pennsylvania, and Wisconsin.

In Michigan, 36% of residents statewide approve of his job performance, and 55% disapprove.  In Pennsylvania, 33% approve of how President Trump is doing in his post, and 52% disapprove.  Among Wisconsin residents, the president’s score is similarly upside down, 33% to 56%.

Those are, by historical standards, pretty weak numbers for a president to have in states he won less than a year ago. But a lot of the coverage has omitted some important context: A lot of the same people who disapprove of the job Trump is doing may have disapproved of him back in November, and voted for him anyway because they preferred him to Hillary Clinton. The exit polling in Wisconsin, for example, found that 56 percent of Trump’s voters had reservations about him and another 23 percent said they were supporting him because they disliked his opponent. Only 35 percent of voters said they had a favorable impression of him–which is not too far from the 33 percent who approve of his job performance.

It’s a jump, then, to assume that Trump has lost a lot of support since the election or that these numbers show that he is in trouble for re-election in 2020. If a lot of voters who have qualms about him have more serious ones about the Democratic nominee, he could still win with numbers like these.

But the opposite mistake would be to assume that since polling generally indicates that Trump would still beat Clinton if the election were held today–see here for a recent example of such a poll–he is in the clear. The Democrats might nominate someone stronger next time.


Ramesh Ponnuru is a senior editor for National Review, a columnist for Bloomberg Opinion, a visiting fellow at the American Enterprise Institute, and a senior fellow at the National Review Institute.